基于文本分析的人物画像技术研究——以知乎大V为例

Research on Character Portrait Technology Based on Text Analysis—Taking the Influencer as an Example

社交媒体高影响力用户集独特的内容、自带话题的商业价值以及高效的流量变现能力于一体。构建高影响力人物画像,可以直观展示群体人员的典型特征,对扩大优秀网络文化的辐射力以及对平台提供精准服务、维系核心用户、监管引导舆情等具有重要作用。我们以知乎大V为例,利用Python获取和处理用户行为数据,从知乎大V多个维度构建用户画像概念模型,用“WordCloud”、“Mapplotlib”以及“Pyecharts”等工具进行可视化,开展实证研究。实验结果表明知乎大V主要分为两大类,以随机选取的领域大V作为典型代表,得到不同高影响力用户的典型人物画像,依据用户及其粉丝的关键群体特征属性,表明用户画像模型可以在社交平台的个性化服务、推荐系统、精准营销等方面带来巨大的应用价值。

Social media high-impact users combine unique content, the commercial value of their own topics, and efficient traffic monetization capabilities. Constructing portraits of high-impact characters can directly display the typical characteristics of group personnel, which plays an important role in expanding the radiating power of excellent network culture, providing accurate services to the platform, maintaining core users, and supervising public opinion. Taking influencer as an example, we use Python to acquire and process user behavior data, construct user portrait conceptual models from multiple dimensions of influencer, use tools such as “WordCloud”, “Mapplotlib”, and “Pyecharts” to visualize and develop Empirical Research. The experimental results show that influencer is mainly divided into two categories. Taking randomly selected domains as typical representatives, we obtain typical portraits of different high-impact users. According to the key group characteristics of users and their fans, it shows that the user portrait model can bring tremendous application value in the aspects of personalized service, recommendation system and precision marketing of social platforms.